The Content Problem

[Dragon45]

Engines are common and reusable. Lots of free and open source engines are available, and if you don't like the FOSS options, just mod a closed engine. For modern games, the place where the most effort is expended is the content. This is a problem for all games, but especially for free or open source efforts, since they tend to have more limited resources. I'm loosely defining content as anything that the user directly sees, including effects, textures, and maps, but especially models, textures, and maps. The "content problem" has three parts.

1) Content is non-reusable. Content is closely associated with branding and trademarking, so its difficult to find legal justification for reuse except by the company that made it. Even if legal justification is found, games that use other games' content are seen as ripoffs, copycats, or clones by players; the exception is games that are in a series, or expansion packs, etc.

2) Content is expensive in terms of time and manpower. It takes a lot of work and iteration to produce high-quality content. Usually, a game's art and content design team will go through multiple design and planning stages, discarding many, many prototypes and having lots of boring meetings to see what to keep and what to discard. A modern commercial game will have at least as many content developers as engine developers.

3) Content quickly gets stale with time. Content that is high quality in 2001 will be considered average in 2002 and out of date by 2003. Engines get dated too, but since they're not user-facing, an older engine with up-to-date content will appear newer than a bleeding-edge engine using out-of-date content. Graphics engines are a possible counterpoint here, but even older graphics engines used high-resolution textures and higher-polygon models can generally make a game seem new.

Some implications:
1) No matter how good a spring mod looks when it starts off, it will become graphically stale. This means that Imperial Winter, even though it looks awesome now, become average or sub-par looking in comparison to the larger industry average. And a year or a year-and-a-half later, it'll be old.

2) No matter how hard WZ and Co. and smoth try, they'll never ever have a finished product that matches the industry best. The best bet they have is matching older graphics. And even the definition of older will change quickly. There is a game of perpetual catch-up going on here, one that resource-bound projects simply cannot overcome.

So this seems somewhat obvious. It is also obvious that we are not playing its mods for their bleeding edge graphics capabilities. "Oh Dragon45, why did you write this?" you may ask.

Well, I'm specifically asking because I'm wondering if anyone reading this has a solid academic (a high-level course or three) or professional experience in either
1) machine learning techniques (rules based systems, advanced statistics, GAs/ANNs) or
2) image processing (i dont suppose we have any photosynth devs here?)
3) information retrieval (SVMs and kernel machines in particular)
4) advanced procedural content generation

I think you can see where I'm going with this. And yes, i am slightly insane. But I just might know what I'm doing.

Regardless, consider this the prelim to a kind of RFC. Feedback appreciated.

-Dragon45

[Guessmyname]

2) No matter how hard WZ and Co. and smoth try, they'll never ever have a finished product that matches the industry best. The best bet they have is matching older graphics. And even the definition of older will change quickly. There is a game of perpetual catch-up going on here, one that resource-bound projects simply cannot overcome.

Wouldn't this be true of all games? Industry or freeware? It's not like the game companies would remake an old game to 'bring it up to standard'. Not unless it's a full-blown remake and they rarely come along for decades

[SerjnDestroy]

(In which mr.newbiedontknownothing posts )

I'm not at all as deeply involved in developing/moding as anyone else here, but I'd like to throw in my 2 fils worth (and hope I do not overstep my outsider bounds)

1) I don't totally agree with that. Highly recognizable things, sure, but it is surprising how much content you can share between projects with minimal remixing without getting the "clone" feel. Filters and batch editing can also go a long way into freshening up and distinguishing one set of content from another http://search.freegamedev.net/projectmap.png (an example of media/content sharing between games). Of course, when you start copying mainstream games, whit a higher recognition factor, things are going to get cloney very quick, but in general, content is shareable in the foss world.

2) Yep, I agree totally. Time, effort and quality control are all very difficult things to manage when its a nonprofit/community project. However, procedural generation is a very underused yet useful tool in this regard, if you like the style and can get your head around it.

3) In my opinion, style > technical specifics. People are still playing /ancient/ games, with all their graphical rustyness, and enjoying it. Sure, that doesn't mean that you should stop improving things, but that doesn't mean that everything done now will be inadequate for the future..

So, yeah, hope my rant has been of some use.. The length of my posts always seem directly proportional to the amount of time I have gone without sleep, heh, the sense/value they make inveresly

[Gnomre]

I think we can all agree that Spring needs more tits

[KDR_11k]

Graphical progress on the PC has come to pretty much a standstill since more and more games are designed for consoles and only ported to the PC. I think it has become less of a moving target, especially with the console market looking like graphics development will slow down there too (namely the Wii vastly outperforming the competition). Additionally there is such a thing as good enough. While e.g. CoH looks awesome and all I don't think it'd have suffered much if the units were made to a lower spec as they're seen from pretty far away and all those details like individual fingers and facial animation on the infantry are only seen in cutscenes.

Of course I don't think any Spring project has graphics that are on par with commercial quality and that's not because of any moving targets, it's because noone here has the skill level of a pro. Maybe in part because Spring's animation system is prehistoric, too.

[Argh]

<shrugs>

Other than the MD5 thing... which appears, at this point, to have a release date of "never"... we have most of the tools.

We can't have normal maps until somebody sticks that GPL'd code I found into Spring, though, so that we can get the tangents on S3Os. That's a major element of the suck, graphically- how in the hell are we supposed to compete with anything modern, when lacking what is a pretty standard part of artists' toolkits nowadays?

That GLSL stuff I did earlier could be re-purposed for a lot of things, and that code was made freely available.

And I was talking to somebody about drawing stuff with the GPU the other day to speed up performance of certain rendering tasks.

Lastly, World Builder exists, as a real, practical (and demonstrably working) repository of content. There aren't any excuses, and pretty much no real need for talk-talk... if you want to make content, make some content, or be... "content" with whatever the people who do things do.

I guess, tbh, that I find this discussion is mainly just useless. We already know what we need, to speed up production times and improve the quality, but we don't have the tools.

[smoth]

Content only becomes stale when you let it. Content only quickly becomes outdated if you are shooting for what is acceptable now. even industry companies pick higher requirements than what is required of what is out now.

Most of the current tech that people drool over are normal maps and shader systems. bleh.

mostly I don't like procedural crap it lends it's self to rather crappy looking stuff.

Dragon, you get 7/10 which isn't bad but you should have said more about how all the current work is already outdated. Also if you are looking for volatility, you should have tried saying shit about pure or gundam to the effect that gundam is just a ta skin and pure is shoddy. After all you get a back to back post from two of the more potentially volatile members of the community but it lacks all nerdrage. Before someone rages, the above paragraph is a JOKE or is it...

[Snipawolf]

TA is shoddy and GundamRTS is a PURE skin!

(gleefully waits for nefarious plan to explode)

[KDR_11k]

Argh, you overrate the importance of Worldbuilder. From what I've seen the results so far are a really ugly version of SimCity. Real cities don't look like that and games have moved way beyond obvious grid designs.

[Argh]

<shrugs>

Do better. Do it as fast.

You can't do either.

[Gota]

I just wanna ask.
It's not that I have some perticuler attraction to ARGH's post and i think they are mostly too long but it seems he does all sorts of stuff like world builder and ads things to the engine etc..also pure is some very nice work even though it s gameplay far form being complete.
to me it seems his work is underrated...

I was wondering why...i think i might have missed some bad things he did and i wanna catch up..or you just dont liek his personality?(what matters is what someone does not what he says and how he says it?)

So what have i missed?

[Pxtl]

Obviously-gridded cities are better than none at all, by a long shot.

@Yan

Argh is a serious modder. Like all modders here, his ego is the size of Cleveland, and he's as sensitive as a puddle of nitroglycerin. He frequently posts algorithmic ideas that are outright wrong and is obviously a bit of a neophyte to serious programming, and gets in constant flamewars with the other modders, etc. However, he's also produced about a hundred times more quality work than most of us, so we love him anyways, and the workflow he's developed for getting a model from concept to decent-looking RTS units gives me wood.

Yay Argh!

[Dragon45]

as long as argh shuts up and produces stuff, everyone loves him. as soon he starts opening his mouth, problems arise.


that being said, has anyone even bothered to read the last four points in my post (particularly about people with the skills/knowledge i was asking for?)

[Ashnal]

From what I've seen of world builder, thats just arghs poor use of the nice tool hes created. The tool itself is great, argh just sucks at designing interesting looking building arrangements. In any case world builder opens up another possibility for content developers.

[Argh]

You've got me there. Brilliant level designer, I ain't.

That said, it's working... and Dragon, you have yet to spit out anything like a serious design proposal, let alone working code, which is why I'm giving you a hard time.

Instead of telling us you want people who are 'serious' and 'qualified'... and giving out acronyms and a few fancy terms for "let's try and emulate Spore"... give us a real proposal to build a real product. Right now, what you're talking about is, well, very vague. Where would content developers come in? Where do the programmers come in? What are the proposed final features?

[smoth]

yeah I know argh, that last volley was what? 5/10 or would you give it a 3?

[Argh]

IDK, about 3.5 on the Gota scale, maybe.

Meh, I'm fucking stressed, people, and it's not really anything about stuff going on here.

I have a bloody meeting tonight for a proposal I'm finishing right now... I'm stressed and I'm just not in the mood to deal with snarky people. Been very, very stressed lately, I'm juggling multiple chainsaws.

[smoth]

fuck the chainsaws use sharks. I'd say pretend they are naked but odds are they might be all dudes, and if it is hot chick... Eh, I was nervous at my interview for my current job. Above all honesty and a positive attitude are important.

[Argh]

Sharks are too slippery right now, I'd just get bit. I hate writing proposals, they're a pain in the arse, even with a template. At least this is an informal meeting, and I probably won't look too bad if it's not entirely perfect

At any rate... give us more details, Dragon. I apologize for being an arse, I assume you really have a idea here, I'm just too stupid to see how you're going to actually achieve it.

[KDR_11k]

<shrugs>

Do better. Do it as fast.

You can't do either.

Um, Argh, "you can't do it better so you cannot criticize it" is just a pure fallacy.

Besides, it doesn't take much of a tool to handle city design on pro level, just something to store the positions of what you placed and spawn it the next time you run the game. Procedural generation has no place in city design like that, the map size is limited and the city has a massive gameplay impact so it should be designed by hand.

What the pro tools do more is stuff like map texture design since almost any game uses splatting or what it's called, tons of smaller textures overlayed and whatnot to create the final map texture. Not gonna happen with smf of course.